Taming of the slew rate measurement

[jackinnj]

Taken from Jim Williams application note #94 "Slew Rate Verification for Wideband Amplifiers" -- I liked the title as I was noodling around.

So the formula goes like this --

SR = 2 * pi * BW * SQRT (2*P*R)

for specifying the "full power" bandwidth, does one mean "that point" in the distortion vs frequency plot where distortion becomes significant, i.e. over 0.1% or 1.0% -- or does it mean that point where the output power is down a certain decibel value from 20kHz?

[andy_c]

Originally posted by jackinnj
Taken from Jim Williams application note #94 "Slew Rate Verification for Wideband Amplifiers" -- I liked the title as I was noodling around.

So the formula goes like this --

SR = 2 * pi * BW * SQRT (2*P*R)

for specifying the "full power" bandwidth, does one mean "that point" in the distortion vs frequency plot where distortion becomes significant, i.e. over 0.1% or 1.0% -- or does it mean that point where the output power is down a certain decibel value from 20kHz?

This sounds like it's related to Mr. Shake-a-spear (or is it Shakespierre?) somehow .

But seriously, for a sine wave v(t)=V_msin(omega*t), the maximum time derivative is omega*V_m (at the zero crossing). In your formula above, SQRT(2*P*R) is just the peak voltage V_m. So the full power bandwidth is the frequency at which the maximum time derivative of the the sine wave output voltage at full rated power is the same as the slew rate.

[thanh]

Hi andy_c! I use LTspice so how can I measure Slewrate?
Thanks!

[andy_c]

Originally posted by thanh
Hi andy_c! I use LTspice so how can I measure Slewrate?
Thanks!

Hi thanh,

Using transient analysis, apply a sine wave input so that the output is just below clipping. Now keep increasing the frequency of the sine wave until the output looks very distorted, like a triangle wave with sort of straight edges instead of curved. Right click on the V(out) expression (or whatever the name of the voltage at the top of the plot is) and in the combo box in the upper right, where it says "attached cursor", choose "1st & 2nd". This will give you two cursors. Drag them to two points on an edge of the triangle wave. At the bottom, where it says "Diff(Cursor2-Cursor1) take the data that says "slope" (which will be in Volts per second) and divide it by 1e6 to get volts per microsecond.

[thanh]

Thanks Andy_C!